Does Trade with Low-Wage Countries Hurt American Workers? · Does Trade with Low-Wage Countries Hurt American Workers? Stephen Golub* T here are gaping disparities in wages and benefits

Has Globalization Created a Borderless World? Janet Ceglowski

3

Does Trade with Low-Wage CountriesHurt American Workers?

Stephen Golub*

There are gaping disparities in wages andbenefits around the world. In 1996, averagehourly earnings of production workers inmanufacturing were $31.50 in West Germany,$17.20 in the United States, $1.51 in Mexico, andless than $0.50 in India and China. How cansuch huge wage differences exist? Are Ameri-can workers’ wages and benefits forced downby competition from low-wage countries? Aretrade barriers the solution? While there aresome genuine problems raised by trading withlow-wage countries, this article will try to showthat popular fears are based on misunderstand-

ing of the causes and effects of wage dispari-ties.

The following quotation, from the conclud-ing article in the September 1996 PhiladelphiaInquirer series “America: Who Stole theDream?” by Donald Barlett and James Steele,forcefully expresses the widely held view thatcompetition from goods produced in low-wagecountries is unfair and detrimental to Ameri-can workers.

"Companies that produce goods in foreigncountries to take advantage of cheap laborshould not be permitted to dictate the wagespaid to American workers."

"A Solution: Impose a tariff or tax ongoods brought into this country equal to the

*Steve Golub is professor and chair, Department of Eco-nomics, Swarthmore College, Swarthmore, PA. When thisarticle was written, he was a visiting scholar in the ResearchDepartment of the Philadelphia Fed.

4 FEDERAL RESERVE BANK OF PHILADELPHIA

BUSINESS REVIEW MARCH/APRIL 1998

wage differential between foreign workersand U.S. workers in the same industry. Thatway competition would be confined to whomakes the best product, not who works forthe least amount of money.

"Thus, if Calvin Klein wants to makesweatshirts in Pakistan, his company wouldbe charged a tariff or tax equal to the differ-ence between the earnings of a Pakistaniworker and a U.S. apparel worker....

"If this or some similar action is not taken,the future is clear. Wages ofAmerican workers will con-tinue to slip, as well as theirstandard of living."

These arguments ignore a fun-damental point: differences inwage rates between countrieslargely reflect differences in laborproductivity (output per hourworked). For example, wagesare low in India because produc-tivity is low. Thus, the costs ofproducing goods are not as dif-ferent across countries as wagerates suggest. Indeed, the UnitedStates as a whole benefits frominternational trade, irrespectiveof the wage levels of its tradingpartners, by specializing in whatwe do well and importing goodsthat are most efficiently pro-duced elsewhere. By increasingefficiency, international trade,like technological change, in-creases the size of the economicpie available to the nation.Granted, international tradedoes adversely affect some in-dustries and individuals, espe-cially in the short run, but thereare more than offsetting benefitsto the rest of the economy. Ratherthan hobbling the efficiency of

the American economy with trade restrictions,it is better to ease the burden on the minorityof Americans who are adversely affected.

MAGNITUDE OF INTERNATIONALDIFFERENCES IN WAGES AND BENEFITS

Labor costs in the industrialized countriesare much higher than those in the developingcountries, although labor costs vary greatlywithin each group, too (Table 1; Figure 1). U.S.manufacturing wages are well below those of

TABLE 1

Indicators of Hourly Labor CostsFor Production Workers in

ManufacturingSelected Countries, 1996a

Labor Costs Labor Costs(in $U.S.) (As a Percent

of U.S.Labor Cost)

United States 17.74 100Canada 16.66 94France 19.34 109Germany 31.87 180Italy 18.08 102Japan 21.04 119United Kingdom 14.19 80

Hong Kong 5.14 29Korea 8.23 46Mexico 1.50 8Singapore 8.32 47Sri Lankab 0.48 3

aLabor costs in other countries are converted to U.S. dollars at themarket exchange rate. Labor costs include wages and fringe benefits.

b 1995

Source: U.S. Bureau of Labor Statistics


5

Germany but above those of the United King-dom. For medium-income countries like Ko-rea, labor compensation levels in manufactur-ing have reached nearly half of those in theUnited States, while low-income countries suchas Sri Lanka, India, and China have labor coststhat are less than 5 percent of U.S. levels.1

THE PRINCIPLES OF COMPARATIVEAND ABSOLUTE ADVANTAGE

Popular discussions confuse the relation-ships between international trade, wages, andlabor productivity. Wages are determined by theoverall productivity of labor (absolute advan-

tage) and are therefore not an independentsource of international competitiveness. Tradepatterns depend on comparative advantage:industry-by-industry differences in productiv-ity across countries. We will first consider thesebasic principles before turning to the evidence.

The important distinction between compara-tive and absolute advantage, first put forth byDavid Ricardo in 1817, is best explained with asimple example (Table 2). With no internationaltrade, the United States demonstrates higherproductivity than Mexico in both industries inthis example, but the productivity ratio isgreater in computer chips (10 to 1) than in shirts(2 to 1).

To produce more shirts, a country must sac-rifice chip output and vice versa, given a lim-ited supply of workers. The number of chipsthat must be given up to produce, say, one moreshirt is what economists call the “opportunity

1Labor costs in manufacturing differ by industry; how-ever, these industry variations are swamped by the overalldifferences in wages between countries. Therefore, it is notmisleading to focus on manufacturing averages.

FIGURE 1

Hourly Labor CompensationOf Production Workers in Manufacturing*

Selected Countries

*Labor costs in other countries are converted to U.S. dollars at the market exchange rate. Labor costs includewages and fringe benefits.

Source: U.S. Bureau of Labor Statistics

Does Trade with Low-Wage Countries Hurt American Workers? Stephen Golub



cost” of a shirt. Since a worker in the UnitedStates can produce 10 chips or two shirts, theopportunity cost of one shirt is five chips. InMexico, since a worker can produce one chipor one shirt, the opportunity cost of one shirt isone chip. Thus, the opportunity cost of shirtsis higher in the United States than in Mexico.Therefore, Mexico has a “comparative advan-tage” in producing shirts, since it has a loweropportunity cost: that is, producing shirts“costs” fewer chips. Similarly, the United Stateshas a comparative advantage in producingchips, since its opportunity cost in that indus-try is lower.

As the example suggests, the determinationof comparative advantage depends only on theratio of productivity in the two industrieswithin each country. For example, if Mexicanproductivity were to double, so that eachworker could produce either two chips or twoshirts, the opportunity cost would be un-changed, and Mexico would retain its compara-tive advantage in producing shirts.

A related concept is that of absolute advan-tage. A country is said to have an absolute ad-vantage in producing a good if a worker in thatcountry can produce more of the good than aworker in the same industry in a different coun-try. In the example above, the United States hasan absolute advantage in producing both chipsand shirts because a U.S. worker could producemore of either good than a Mexican worker.

Despite this absolute advantage, however,

the total output of the world economy—and thestandard of living in each country—will behigher if U.S. workers produce more of thoseitems in which they have a comparative advan-tage and Mexican workers do the same, and thetwo countries trade. In general, absolute advan-tage determines the overall level of wages ineach country, and comparative advantage de-termines trade patterns.

To put this concept simply, let’s supposewages in the United States are five times thosein Mexico—as they were before Mexico’s cur-rency crisis in 1994—in both the shirt industryand the chip industry.2 Since U.S. workers canproduce 10 times as many chips as their coun-terparts in Mexico, but their wages are only fivetimes as high, the United States will have lowerlabor costs per chip. Similarly, since U.S. work-ers produce only twice as many shirts as Mexi-can workers, but their wages are five times ashigh, the United States will have higher laborcosts per shirt. So, ideally, Mexico should pro-duce more shirts, the United States should pro-duce more chips, and the two countries shouldtrade. Such a transaction produces more goodsat lower cost because it allows each country toproduce more goods in the industry in whichit has a comparative advantage.

Both countries’ living standards will increasefrom trading according to comparative advan-tage because the resulting world pattern of pro-duction is more efficient than if each countryproduces only for its own market. The UnitedStates can obtain shirts more cheaply fromMexico than by producing shirts itself, payingfor these shirt imports with chip exports. In-ternational trade does not cost U.S. jobs, but itdoes change the industry mix of U.S. outputand employment. American production of

2Wages for workers with similar characteristics will bethe same in different industries within a country if the la-bor market is competitive and workers can freely move be-tween industries.

TABLE 2

Output per WorkerPer Hour

ComputerChips Shirts

United States 10 2

Mexico 1 1


7

That Thing Venture Capitalists Do Mitchell Berlin

chips will expand while shirt production con-tracts, resulting in corresponding shifts in la-bor demand. The reverse happens in Mexico.

There are two qualifications to this charac-terization of the benefits of trade. First, relocat-ing workers between the shirt and chip indus-tries may be difficult in the short run, resultingin some unemployment of former shirt work-ers in the United States. Second, this kind oftrade may reduce unskilled workers’ real wagesin the United States, even after workers are re-located, if the chip industry employs a higherratio of skilled to unskilled workers than theshirt industry. In the United States, as chip pro-duction expands and shirt production falls, thedemand for skilled labor rises, while the de-mand for unskilled labor declines. As discussedlater, however, the proper response to these dis-tribution effects is not to restrict trade but toease the transition by retraining displacedworkers.

These days, international trade, which is of-ten conducted by multinational corporations,increasingly takes the form of trade in interme-diate products, but the basic gains from tradeare unaffected. American companies locate thesimpler parts of their production processes indeveloping countries, while the more sophisti-cated components are produced at home. Forexample, 21 months after the North AmericanFree Trade Agreement (NAFTA) went into ef-fect, the Key Tronic company, a large manufac-turer of computer keyboards, laid off 277 work-ers in Spokane, Washington, as it relocatedsome of its assembly jobs to a plant in CuidadJuarez, Mexico. But Key Tronic’s chief financialofficer reported that employment in its Spokaneplants actually increased overall because manyof the components used in the keyboards aremade in Washington, and the lower costs ofassembly in Mexico enabled the company tolower prices and increase sales.3

Other studies show that economic integra-tion with Mexico has entailed a boom in manu-facturing production in U.S. cities along the

border because Mexican factories specialize inassembly, which makes intensive use of un-skilled labor, while border regions in the UnitedStates specialize in high-technology tasks suchas production of components and product de-sign.4 This international division of labor fol-lows the principle of comparative advantage.The United States is likely to have an absoluteadvantage in all stages of the production pro-cess, because American workers are, on aver-age, more skilled and educated than those indeveloping countries, and infrastructure in theUnited States is superior. But the United States’advantage in terms of efficiency is likely to begreatest in high-technology production pro-cesses, for which a highly skilled work force iscritical. The United States gains from the in-crease in efficiency resulting from the globaldivision of labor, just as in the simple chip/shirtexample.5

In fact, the chip/shirt example illustrates akey point: low wages most likely reflect lowproductivity. Furthermore, if low wages wereall that mattered in international trade, coun-tries with rock-bottom labor costs, such asBangladesh, Bolivia, and Burundi, would bemajor exporters. Yet, popular concern often fo-cuses on countries such as Mexico and SouthKorea—countries with wages well above thosein Africa and South Asia. Clearly, labor produc-tivity matters, too.

Some people worry that as low-wage coun-

3“NAFTA Tradeoff: Some Jobs Lost, Others Gained,”New York Times, October 7, 1995.

4See the article by Gordon Hanson.

5Robert Feenstra and Gordon Hanson provide a theo-retical analysis of this form of comparative advantage. Onedifference between their results and the textbook analysisis that skilled labor reaps the gains from trade in both theUnited States and the low-wage country. This result is con-sistent with some evidence that the gap between the wagesof skilled and unskilled workers is widening in develop-ing countries, just as it is in developed countries.




tries acquire technology and capital, their pro-ductivity will rise, giving them a competitiveedge. But there are two reasons not to be con-cerned about this. First, as productivity in acountry rises, wages tend to rise as well, so thecompetitive edge lessens. Second, other factors,such as low levels of human capital (knowledgeand skills) as well as poor public infrastructureand transportation services, tend to hold downproductivity in low-wage countries, even whenthey acquire new physical capital (computersand factories). Except for products and produc-tion processes that require large amounts of un-skilled labor, these factors offset the appeal oflow wages for companies considering relocat-ing their production to poor countries.

In addition, developing countries may havehigher costs of other inputs, such as capital,energy, and raw materials. Prices of these in-puts are more likely than wage rates to be simi-lar across all countries, because, unlike labor,nonlabor inputs can be moved across bordersin response to international price differences.Nonetheless, capital, energy, and raw materialcosts per unit of output could be higher in devel-oping countries if these countries use nonlaborinputs less efficiently than developed countries.

In summary, both developed and develop-ing countries can benefit from specializing inwhat each produces relatively efficiently, re-gardless of the overall level of labor costs, be-cause low wages do not necessarily mean lowproduction costs across the board. Low wagesmay be offset by either low labor productivityor higher costs of nonlabor inputs such as capi-tal, energy, and raw materials. Only in low-skillindustries and unsophisticated production pro-cesses are developing countries likely to havelower average costs of production and, hence,a comparative advantage.6

WAGES, PRODUCTIVITY,AND TRADE: EVIDENCE

Wages and labor productivity are related(Figure 2).7 For example, in 1990 wages in Ma-laysia were 10 percent of wages in the UnitedStates. But Malaysian labor productivity wasalso about 10 percent of the U.S. level in 1990.This means that unit labor costs (the ratio ofwages to productivity) were approximately thesame in Malaysia and the United States becausethe difference in productivity almost exactlyoffset the difference in wages between the twocountries. In this case, companies have no in-centive to shift production from the UnitedStates to Malaysia.

In general, international differences in unitlabor costs are much smaller than differencesin wage rates because the huge internationaldisparities in wages mostly reflect equally largedifferences in productivity. In fact, these calcu-lations indicate that, in 1990, unit labor costs inthe Philippines and India were actually higherthan those of the United States, that is, the pro-ductivity difference was even bigger than thewage difference.

Some disparities between wages and pro-ductivity are to be expected for several reasons.

6China’s efforts to develop an aircraft industry are of-ten presented as a counterexample. But China’s exportsconsist overwhelmingly of low-technology items such asclothing, shoes, and toys.

7Productivity is calculated as real value-added per em-ployee. Value-added is the value of output minus the costsof raw materials and other intermediate inputs. Wages aredefined as earnings per employee. Earnings here includeall direct payments, including maternity and vacation payand payment in kind, but exclude employer contributionsto social insurance funds, as data on the latter are not avail-able for most developing countries. The exclusion of socialinsurance costs is likely to overstate relative labor costs indeveloping countries, but only to a minor extent. Direct payis still, by far, the larger part of compensation, accountingfor 70 to 90 percent of total labor compensation even forthe United States and other rich countries. The ratio of em-ployer-paid benefits to total labor costs is not that muchhigher in developed countries compared with the few de-veloping countries for which this information is available.For details on sources and methods of the calculations ofwages and productivity, see my 1995 article.


9


First, differences in wages sometimes reflecttemporary exchange-rate movements, whichmay have little effect on long-term businessdecisions about the location of production. Forexample, the appreciation of the dollar againstthe mark and the yen in the early 1980s sharplylowered German and Japanese wages measuredin U.S. dollars (see Figure 1). The depreciationof the dollar in the late 1980s and early 1990s,however, led to a large increase in German andJapanese wages expressed in U.S. dollars.8 Sec-ond, as noted above, some differences in unitlabor costs may be offset by nonlabor costs, solow unit labor costs do not necessarily imply acompetitive advantage. Third, the available

measures of labor costs and productivity are notalways fully reliable and comparable, especiallyfor developing countries. Despite these quali-fications, a fairly close correlation between la-

8German unit labor costs in the mid-1990s reached lev-els nearly double those of the United States, as Germanlabor compensation rose well above U.S. wages and Ger-man productivity remained at about 80 percent of the U.S.level. Germany’s high unemployment may reflect, in part,the relatively high level of German labor costs. The depre-ciation of the mark in 1996-97 has partially restoredGermany’s cost competitiveness. A similar description ap-plies to recent Japanese unit labor costs, but to a lesser ex-tent.

FIGURE 2

Labor Productivity, Wages, and Unit Labor Costsin Developing Countries, Relative to the United States

Source: Golub (1995)




bor costs and labor productivity is observedacross countries.

Wages and labor productivity also move to-gether over time for individual countries. Forexample, Korea experienced both high wagegrowth and high productivity growth in manu-facturing over 1970-95, compared with theUnited States (Figure 3). In 1970, Korean wageswere 8 percent of U.S. wages, while Korean pro-ductivity was 14 percent of U.S. productivity.By 1995, Korean productivity had reached 69percent of the U.S. level, while Korean wagesgrew to 48 percent of American wages. Notethat U.S. manufacturing productivity andwages grew steadily over this period, so Fig-ure 3 indicates very strong growth in Koreanwages and productivity. Korean workers havegreatly benefited from Korea’s phenomenaleconomic growth.

In Mexico, wages and productivity movedclosely together until the outbreak of the debtcrisis in 1982. This crisis led to policies of ex-treme austerity and steep depreciation of thepeso to enable Mexico to service its foreign debtand, in turn, caused a steep decline in the dol-

FIGURE 3

Wages and Labor Productivity,Expressed as a Ratio of U.S.

Wages and ProductivityKOREA

lar value of Mexican wages (Figure 4). Mexicanwages recovered relative to productivity after1986, but fell back after 1994. This decline inMexican wages and unit labor costs in 1994-95and the subsequent shift of the Mexican tradebalance from deficit to surplus are often inap-propriately cited by U.S. opponents of the NorthAmerican Free Trade Agreement as vindicationof their views that NAFTA would create a “largesucking sound” of jobs being siphoned off toMexico. As in the early 1980s, the drop in Mexi-can wages after 1994 reflects the collapse of thepeso and deep recession in Mexico. Indeed,manufacturing employment in Mexico droppednearly 10 percent in 1995. As the Mexicaneconomy recovers from the crisis, its wages andunit labor costs are likely to increase, as theydid from 1987 to 1991.

The volume of trade is also inconsistent withfears about the competitiveness of low-wagecountries (Table 3). Many developing countries’exports of manufactures to the industrial coun-tries have increased rapidly, but the majorityof these developing countries continue to runtrade deficits in manufactures, as their imports

FIGURE 4

Wages and Labor Productivity,Expressed as a Ratio of U.S.

Wages and ProductivityMEXICO


11


TABLE 3

Developing Countries’ Trade in Manufactured Goodswith All Industrial Countries, Selected Years, as a per-

cent of Developing Countries’ GDP

Exports to Imports From Trade BalanceIndustrial Countries Industrial Countries

Brazil 1970 0.3 2.8 -2.51980 1.1 2.2 -1.11990 2.2 1.5 0.61995 1.7 3.1 -1.4

India 1980 1.1 1.8 -0.71990 2.1 2.2 -0.11995 3.8 3.3 0.5

Korea 1970 6.1 9.8 -3.71980 14.3 11.4 2.91990 15.2 11.6 3.51995 12.3 13.9 -1.6

Malaysia 1970 8.0 13.6 -5.51980 9.3 19.1 -9.81990 19.1 31.3 -12.21994 33.2 45.0 -11.8

Mexico 1970 0.8 4.3 -3.41980 0.7 5.7 -5.11990 3.7 6.8 -3.11995 19.3 16.8 2.5

Thailand 1970 1.3 9.6 -8.31980 4.2 9.4 -5.21990 10.7 17.3 -6.61995 12.8 21.9 -9.1

Sources: United Nations, International Monetary Fund.




have grown nearly as much. For many of thesedeveloping countries, two-way manufacturingtrade with the industrial countries is now quitelarge in relation to their gross domestic prod-uct (Brazil and India are exceptions). Trade inmanufactures is, on the whole, much more im-portant for the developing countries than forthe developed countries, as measured by shareof respective GDP.

In summary, wage differences do mostly re-flect productivity differences. Macroeconomicshocks and exchange-rate fluctuations, how-ever, can entail large discrepancies for severalyears.

INTERNATIONAL TRADEAND THE U.S. LABOR MARKET

U.S. Employment Performance. Critics ar-gue that the overall U.S. trade deficit and thedeficits with particular developing countriessuch as China and Mexico reduce the numberof jobs in the United States. As evidence, theyoften cite the decline of manufacturing employ-ment. They claim that other countries, such asJapan and those in Western Europe, have lessopen markets and consequently do not runtrade deficits like the United States. But thesearguments ignore the fact that overall U.S. em-ployment growth has been extraordinarily im-pressive, far outpacing that of Europe and Ja-pan. Indeed, there has been much discussionin these countries about how to emulate U.S.employment performance. In 1997, the U.S.unemployment rate fell below 5 percent, itslowest level since the early 1970s. In recentyears, the labor force and employment haveincreased more rapidly than the population ofworking age: 4 million workers were added in1996 and the first half of 1997 alone. The NewYork Times reported recently that the demandfor labor is so strong that “companies are re-cruiting among those ignored in the past: moth-ers at home with their children, older men whohad retired or been laid off, students, immi-grants, people with criminal records. State offi-

cials [in Louisville, Kentucky] who help formerprisoners get jobs say companies now rejectfewer convicted felons.”9

Therefore, while the U.S. trade deficits dodisplace some workers, any associated joblosses have been more than offset by overall jobcreation. In fact, the causation runs in the re-verse direction: the strength of the U.S.economy, which manifests itself in employmentgrowth, is an important cause of the overall U.S.trade deficit, since imports rise with incomes.Recessions in Japan, Europe, and LatinAmerica, meanwhile, have held down U.S. ex-ports.

Even in manufacturing, international tradehas had a secondary role in affecting employ-ment trends. In 1994, manufacturing accountedfor 16 percent of all U.S. jobs, down from 26percent in 1970. A recent study found that theU.S. trade deficit accounted for only one tenthof this decline; the remainder is mostly due tothe difference in productivity growth betweenmanufacturing and the service sector.10 Asmanufacturing productivity increases, fewerworkers are needed to produce a rising volumeof output, and the released workers shift to theservice sector. Much the same occurred in agri-culture earlier in the century. Technologicalchange and capital investment lowered theshare of employment in agriculture from 44percent in 1900 to 3 percent today. This processwas undoubtedly painful for many displacedworkers, but few today would consider revers-ing the clock on the gains in standard-of-livingafforded by the growth in agricultural produc-tivity.

Nor is it true that the overall “quality” ofjobs has declined as the quantity has increased.

9“Jobs Opening Faster Than They Can Be Filled,” NewYork Times, July 10, 1997.

10See the article by Robert Rowthorn and RamanaRamaswamy.


13


Careful studies show a mixed picture. Jobgrowth has been strong in high-paying as wellas low-paying occupations, as industries haveshifted the occupational mix of their employ-ees. Between 1983 and 1994, jobs in manage-rial, professional, and technical occupationsgrew more rapidly than overall U.S. employ-ment.11 Once again, this does not deny thatsome workers have suffered because of job dis-location and wage declines, sometimes causedby competition from imports. The overall per-formance of the labor market, however, is atvariance with the popular view that interna-tional trade is devastating American labor.

Wage Inequality. Increased inequality ofwages has been one of the most salient featuresof the American labor market in recent decades.While average family income has increased, thegap between higher-paid and lower-paid work-ers has widened sharply.12 Much of the increasein wage inequality reflects a greater demandfor skilled labor, as evidenced by a large increasein the wages of college graduates relative to thewages of workers without a college education.While increased wage inequality is not neces-sarily a bad thing in itself, as it may reflect amore competitive and discerning labor market,the plight of those at the lower end of the in-come distribution is a source of concern. Thequestion here is the role international trade isplaying.

As suggested by the Mexico shirt/ U.S. com-puter chip trade example, international tradewith poor countries can be expected to increasethe relative demand for skilled labor in theUnited States, since the United States expandsproduction in industries that make intensive

use of skilled labor and it imports goods cre-ated largely by unskilled labor. Such trade maycause not just a widening in the wage gap be-tween skilled and unskilled labor but also anabsolute decline in the real income of unskilledworkers. Also, the widening wage inequalityhas coincided with an increase in internationaltrade with low-wage countries, suggesting apossible connection.

Although there may be a connection betweenincreased trade and income inequality, manystudies conclude that international trade withlow-wage countries has played, at most, a sec-ondary role in increasing income inequality. Asa recent survey of the literature concludes,“Nearly all of this research finds only a modesteffect of international trade on wages and in-come inequality.”13 The small effect of trade onwage inequality in the United States is not sosurprising when one considers the small sizeof such trade. Although imports of manufac-tured goods from developing countries haveexpanded rapidly, in 1995 they still amountedto only 3 to 4 percent of U.S. gross domesticproduct (GDP) and 7 percent of the value ofmanufacturing production. More than half ofU.S. imports of manufactured goods still comefrom other industrialized countries, some ofwhich have higher wages than the United States(see Table 1). Most economists think that tech-nological change, which has increased demandfor workers with higher skills, is mainly respon-sible for the rise in the demand for skilled ratherthan unskilled labor and the resulting increasein wage inequality. Many economists believethat advances in information technology, suchas computers and telecommunications, are atthe heart of the changes affecting the U.S.economy.

In the case of technological change, the ben-efits to the overall standard of living outweigh11See the study published by the Committee on Eco-

nomic Development.

12See Peter Gottschalk’s article for a summary of the factsand other articles in the same issue of the Journal of Eco-nomic Perspectives for further discussion.

13See the article by Matthew Slaughter and PhillipSwagel.




the associated dislocations to those whose skillsbecome obsolete. The economic effects of inter-national trade are similar: trade and new tech-nology both raise the general standard of liv-ing while hurting those whose occupationalskills are devalued. Why accept technologicalchange while restricting trade? Many peoplerecognize that new technology entails a shift inthe composition of jobs rather than a net loss ofjobs but fail to understand that the same is truefor international trade. But, as discussed previ-ously, by specializing according to compara-tive advantage, countries increase their produc-tive efficiency with little net effect on job cre-ation.

Although technological change is far moreimportant than international trade as a causeof wage inequality, trade does adversely affectsome workers. Rather than restrict trade, theUnited States should offer a social safety netand retraining, which are better ways of help-ing displaced workers. That way, society can

reap the gains from trade and share them moreequally.

CONCLUSIONSTrade between the United States and low-

wage countries benefits most people in bothplaces, irrespective of wage differences. Differ-ences in wages largely reflect differences in la-bor productivity and are not a form of unfaircompetition. Developing countries tend to spe-cialize in products created mostly by unskilledlabor while the United States specializes in moresophisticated goods. Some unskilled workersin the United States are adversely affected bysuch trade, although factors other than tradeare more important in accounting for increasesin wage inequality. In any event, restrictingtrade is an inferior solution—it is better to helpdisplaced workers adjust rather than deny so-ciety the gains from specialization according tocomparative advantage.


15


REFERENCES

Barlett, Donald L., and James B. Steele. America: Who Stole the Dream? Kansas City: Andrews andMcNeel, 1996. Reprinted from Philadelphia Inquirer series, September 1996.

Committee on Economic Development. American Workers and Economic Change. CED: Washington,1996.

Feenstra, Robert C., and Gordon H. Hanson. “Foreign Investment, Outsourcing and Relative Wages,”in Robert C. Feenstra and Gene M. Grossman, eds., Political Economy of Trade Policy: Essays inHonor of Jagdish Bhagwati. Cambridge: MIT Press, 1995.

Golub, Stephen S. “Comparative and Absolute Advantage in the Asia-Pacific Region,” Pacific BasinWorking Paper Series, Federal Reserve Bank of San Francisco, No. PB95-09, October 1995.

Gottschalk, Peter. “Inequality, Income Growth, and Mobility: The Basic Facts,” Journal of EconomicPerspectives 11, Spring 1997.

Hanson, Gordon H. “Economic Integration, Intraindustry Trade, and Frontier Regions,” EuropeanEconomic Review 40, April 1996.

Rowthorn, Robert, and Ramana Ramaswamy. “Deindustrialization: Causes and Implications,” IMFWorking Paper WP97/42, April 1997, forthcoming in International Monetary Fund Staff Studies forthe World Economic Outlook.

Slaughter, Matthew J., and Phillip Swagel. “The Effect of Globalization on Wages in the AdvancedEconomies,” IMF Working Paper WP97/43, April 1997, forthcoming in International MonetaryFund Staff Studies for the World Economic Outlook.



17

Has Globalization Created aBorderless World?

Janet Ceglowski*

The newest buzzword in the popular busi-ness press is globalization, a word that evokesimages of a world in which goods, services,capital, and information flow across seamlessnational borders. In this world, the choices overwhere to produce, shop, invest, and save areno longer confined within national borders buthave taken on a decidedly global orientation.Some analysts speculate that globalization hasblurred the economic distinctions between

countries, creating a “borderless world” inwhich economic decisions are made withoutreference to national boundaries. For instance,in describing the sphere in which the major in-dustrial economies operate, Kenichi Ohmaeasserts that “national borders have effectivelydisappeared and, along with them, the eco-nomic logic that made them useful lines of de-marcation in the first place.”

The view that national borders have becomeeconomically meaningless is controversial. But,if correct, it has potentially important implica-tions for the world’s economies and theirpolicymakers. One current concern is that, byenhancing access to the labor resources andproducts of low-wage countries, globalization

*Janet Ceglowski is an associate professor, Department ofEconomics, Bryn Mawr College, Bryn Mawr, PA. When thisarticle was written, she was a visiting scholar in the Re-search Department of the Philadelphia Fed.



could already be stunting workers’ living stan-dards in relatively high-wage countries like theUnited States.1 A truly borderless world wouldplace great limits on the ability both to confinethe effects of domestic economic policy withinnational borders and to insulate countries fromforeign economic shocks. In such a world, fi-nancial capital, production activities, and evenworkers could move in response to better op-portunities elsewhere in the world almost aseasily as they could within a given country,thereby undermining efforts to maintain eco-nomic or financial conditions at home that di-verge substantially from those abroad.

The overall level of international economicactivity has escalated in recent years, spurredby a variety of factors ranging from innovationsin information technology to efforts by nationalgovernments to liberalize and deregulate mar-kets. The result has been an impressive expan-sion in world trade, investment in overseasoperations, and international flows of financialcapital. Casual observation suggests that inter-national economic developments are attractinggreater attention from policymakers, produc-ers, and even individuals in their roles as work-ers and consumers. Both the growth in inter-national economic activity and heightened pub-lic awareness are indications of strengtheningeconomic ties between countries. The UnitedStates has participated in this trend and, bymost measures, is considerably more open to-day than it was even 25 years ago.

Does all this mean that national borders nolonger matter for economic decisions? This ar-ticle assesses the relevance of the “borderlessworld” view for U.S. product markets. Al-though the U.S. economy has become more

open, recent research finds that national bor-ders continue to affect U.S. trade flows andproduct prices. In fact, the estimates of theborder’s effects are substantial. A number offactors could be responsible for this finding,including government-imposed barriers totrade, fluctuations in exchange rates, and a va-riety of noneconomic factors such as nationalhistorical and cultural ties. Even in the currentenvironment of global and regional trade liber-alization, there is little reason to expect that theinfluence of these factors on U.S. product mar-kets is about to disappear.

GLOBAL AND REGIONAL INTEGRATION:EVIDENCE FOR U.S. PRODUCT MARKETS

National economies are linked through tradein goods and services, cross-border flows of fi-nancial assets, and labor migration. Interna-tional economic integration is the process bywhich reducing barriers between nationaleconomies strengthens these ties. In the eco-nomics literature, integration traditionally hasbeen associated with explicit government ac-tions to lower tariffs and other artificial barri-ers to the international movement of goods,services, and inputs. Recent advances in com-munication and information technologies havealso promoted economic integration by enhanc-ing knowledge of and access to foreign consum-ers and products. Both trade liberalization andadvances in communication and informationcontinue to be operative factors in the U.S.economy.

Have U.S. product markets become moreintegrated with the world economy as a result?2

1This is, itself, a hotly debated issue among economists.The debate centers on the impact of trade on jobs, wages,and income distribution. See, for instance, the article byPaul Krugman and Robert Z. Lawrence and the sympo-sium papers in the Journal of Economic Perspectives, 9 (Sum-mer, 1995), pp. 15-80.

2 While this paper is concerned with the economic inte-gration of U.S. markets for goods and services, the termcan also be applied to markets for inputs like labor andfinancial capital. By and large, labor market integration islimited by government-imposed barriers to internationalmigration. In contrast, financial capital is perceived ashighly mobile internationally. That view is supported by


19

One common approach to quantifying thestrength of an economy’s ties with the rest ofthe world is to measure the share of its economic

activity made up of exchanges with other coun-tries. A larger share is indicative of a more“open” economy, one with stronger links to theworld economy. According to this measure,markets for goods in the United States havebecome more open. Measured relative to grossdomestic product (GDP), merchandise trademore than doubled between 1970-71 and 1995-96 as a result of significant growth in both ex-ports and imports (Table 1). Much of that gainoccurred in the 1970s, so that by 1980-81 mer-chandise trade was 16.5 percent of GDP. Theexpansion in U.S. trade resumed in the 1990s,albeit at a somewhat slower pace. Though

TABLE 1

U.S. Trade in Goods and Services Relative to U.S. GDP(annual averages; in percent)

1970-71 1980-81 1990-91 1995-96

Merchandise,excluding military 8.0 16.5 15.4 18.5

of which: exports 4.0 7.8 6.9 8.0imports 4.0 8.7 8.5 10.5

Private services 1.8 2.5 4.2 4.8


Merchandise & services 9.8 19.0 19.6 23.3


Notes: The totals are the sums of the individual percentages for exports and imports. Private services trade iscalculated as total services trade minus transfers under U.S. military sales contracts, direct defense expenditures,and U.S. government miscellaneous services.Source: Author’s calculations based on data from Bureau of Economic Analysis

the fact that, at least for the major currencies, interest ratedifferentials between identical offshore and domestic as-sets are insignificant. But as Martin Feldstein observes, thismerely reveals that financial capital can and does moveacross national borders. In fact, despite the substantialholdings of foreign stocks and bonds, recent research indi-cates that investors exhibit a strong home bias in their in-vestment portfolios (see the articles by Kenneth French andJames Poterba; and Linda Tesar and Ingrid Werner). Theimplication is that international capital markets are not fullyintegrated.



smaller in value than goods trade, services tradehas grown even faster: measured relative toGDP, it has nearly tripled since 1970-71. To-gether, exports and imports of goods and ser-vices have expanded from under 10 percent ofGDP in 1970-71 to over 23 percent in 1995-96.3

Do similar measures show evidence of grow-ing regional integration? Recent trade agree-

ments between the United States, Canada, andMexico have created a tri-national free tradearea; the Canada-United States Free TradeAgreement (CUSFTA) liberalized trade be-tween the United States and Canada in 1989 andthe North American Free Trade Agreement(NAFTA) extended the free trade area to Mexicoin 1994. As a result, numerous formal barriersto trade and investment between the threecountries have been or will be eliminated. Thereduction in economic barriers should promotegreater integration of the three economies. Infact, merchandise trade with Canada andMexico grew from 2.3 percent of U.S. GDP in1970-71 to 5.4 percent in 1995-96 (Table 2). Someof that growth predates the creation of theNorth American free trade area, suggesting anongoing process of economic integration be-tween the United States and the two otherNAFTA countries. However, the recent trade

TABLE 2

U.S. Trade with Canada and Mexico Relative to U.S. GDP(annual averages; in percent)

1970-71 1980-81 1990-91 1995-96

Merchandise 2.3 3.9 4.1 5.4

of which: Canada 2.0 2.9 3.0 3.8Mexico 0.3 1.0 1.1 1.6

Private services NA NA 0.71 0.69

of which: Canada 0.30 0.28 0.45 0.44Mexico NA NA 0.26 0.25

Notes: The individual percentages for Canada and Mexico represent the ratios of the sum of exports and importsto U.S. GDP. The totals for merchandise and services are the sums of the individual percentages for Canada andMexico.Source: Author’s calculations based on data from Bureau of Economic Analysis

3It could be argued that trade statistics underestimatethe extent of product market integration because they donot fully account for the contributions of companies’ over-seas operations. For example, foreign companies have in-vested heavily in U.S. production facilities over the last 15years or so. The result has been a significant rise in the levelof economic activity of foreign companies operating in theUnited States. In fact, the Bureau of Economic Analysis es-timates that the output of U.S. affiliates of foreign compa-nies has grown faster than total U.S. output; as a share ofgross output originating in private industries, it has in-creased from 2.3 percent in 1977 to 6 percent in 1995.


21

agreements could have played a part in the sig-nificant gain since 1990-91. They might also bea factor in the sustained rise in the share of pri-vate services trade with Canada.

IS THE U.S. BORDER IRRELEVANT?The preceding analysis indicates that U.S.

product markets have become more integratedwith global markets. There is some indicationof the same phenomenon at the regional level.But evidence of greater economic integrationis not the same as evidence that national bor-ders no longer matter for the worldwide distri-bution of goods and services. Although thisdistinction may appear to be simply a matterof degree, it is important. In a truly borderlessworld, the strength of the economic ties be-tween markets would not depend on whetherthey are located in the same country. In par-ticular, consumers and producers within agiven country would not trade more amongthemselves simply because of shared national-ity. In the language of economic integration,borderless product markets would be tanta-mount to complete integration.

Do borders still matter for U.S. product mar-kets? The border between the United States andCanada is the most likely place to find evidencethat they don’t. Complete economic integrationrequires that there be no trade barriers betweencountries. Therefore, the strongest evidence ofborderless product markets should be foundamong countries that have largely eliminatedbarriers to trade between them. The UnitedStates and Canada are clear candidates: not onlyhave CUSFTA and NAFTA eliminated numer-ous barriers to bilateral trade but, for manygoods, tariffs and other formal trade barriersbetween the United States and Canada werelow or nonexistent well before the recent tradeagreements.

Several other features of the two countriesfavor the development of strong bilateral eco-nomic links. Geographic proximity is one suchfeature. Greater distances between markets

mean larger costs of transporting goods andservices between them, encumbering trade andthe development of close economic ties.4 Butthe United States and Canada share a long bor-der, much of which is easily negotiated by landor water. Moreover, some Canadian cities arecloser to urban centers in the United States thanthey are to other major Canadian cities. Indeed,over three-fourths of Canada’s population liveswithin 100 miles of the U.S. border. The near-ness of the two countries extends beyond merephysical proximity: Canada and the UnitedStates share a number of social, political, andcultural traditions, and a majority of people inboth countries speak the same language. Boththe geographic proximity and cultural similari-ties of the two countries are propitious for bi-lateral trade and other cross-border economicactivities.

In fact, Canada and the United States havelong been major destinations for each other’sproducts and foreign investment. They cur-rently exchange close to $1 billion in goods andservices each day, making theirs among theworld’s largest bilateral trade flows. But howare we to gauge whether this cross-border eco-nomic activity is evidence that the U.S.-Canadaborder no longer matters? One approach wouldbe to evaluate the economic ties between a Ca-nadian market (say, Toronto) and a U.S. market(say, Philadelphia). The strength of the ties be-tween any such pair of markets could dependon a number of factors, including the geo-graphic distance between them and the com-position and sizes of their respective economies.But if economic activity were unaffected by thepolitical border between Canada and the UnitedStates, the strength of the ties would not de-pend on the fact that the two markets are lo-

4James Rauch argues that for all but a few relatively stan-dardized products such as those traded in organized glo-bal markets, greater distances can also raise the costs oflocating appropriate sellers or buyers in foreign markets.



cated in different countries. Evidence to thecontrary would imply that the border doesmatter and that the two economies cannot becharacterized as completely integrated.

Recent research finds that the relatively in-nocuous U.S.-Canada border has significanteconomic effects. The evidence is twofold. First,studies of Canadian merchandise trade revealthat the average Canadian province tradesmuch more with other Canadian provinces thanwith U.S. states of similar economic size andgeographic distance.5 Ontario, for instance, isroughly equidistant from British Columbia andthe state of Washington. Yet in 1990, it tradedover seven times more with British Columbiathan with Washington, despite the fact thatWashington’s economy was almost twice thesize of British Columbia’s. This suggests signifi-cant home bias in Canadian merchandise tradevis-a-vis the United States.

Second, evidence also comes from compari-sons of consumer prices in the United Statesand Canada. If the U.S.-Canada border wereeconomically irrelevant, there would be nolarge, persistent differences between the pricesof identical products in Canadian and U.S.markets, once they were expressed in terms ofthe same currency. As every consumer has ex-perienced firsthand, price differentials for thesame good can and do exist at any single pointin time. According to economic theory, how-ever, the actions of buyers in search of low pricesand sellers in pursuit of profits should mini-mize these price differences over time. Econo-

mists acknowledge that this process can take aconsiderable amount of time. They also recog-nize that prices of similar products in differentlocations may not be exactly equalized, owingto such factors as the cost of transporting theproducts between locations. When markets areintegrated, however, the forces of competitionshould ensure that such prices move in paral-lel with one another over the long run. Yet re-cent research finds little evidence of such a cor-respondence between the U.S. dollar prices ofconsumer goods in U.S. and Canadian markets,even in the long run.6

The empirical evidence clearly indicates thatthe border has economic effects—that is, theborder “matters”—for product markets in theUnited States and Canada. This conclusion maynot be terribly surprising. After all, the free tradearrangement between Canada and the UnitedStates stops far short of establishing an eco-nomic union. A more interesting issue concernsthe magnitude of the border’s effects. That is, ifthe border matters, does it matter a lot? Theanswer appears to be yes. By one estimate, aCanadian province engages in 20 times moremerchandise trade with another Canadianprovince than with an equidistant U.S. state ofcomparable economic size.7 Preliminary evi-

5See the papers by John McCallum and John Helliwell.It is possible that the effects attributed to the border actu-ally derive from differences in the composition of state andprovincial production. That is, interprovincial trade couldexceed trade between Canadian provinces and U.S. statesnot because of the border, but simply because the prov-inces can obtain more of what they want from other prov-inces. However, when McCallum explicitly controls for thispossibility, he finds that it does not account for the largeeffect of the border on provincial trade patterns.

6John Rogers and Michael Jenkins analyze ratios of U.S.prices to Canadian prices (both expressed in U.S. dollars)for various categories of consumer products. In construct-ing the ratios, they carefully pair U.S. consumer productswith similar Canadian products to ensure that they are com-paring the prices of like goods. Even so, they fail to findevidence of a stable, long-run relationship between mostproduct pairs. In a related study, Charles Engel comparesthe variability of price ratios for pairs of consumer prod-ucts in the United States and Canada. He reports that thevariation in the dollar price ratio of similar Canadian andU.S. consumer products is typically much larger than thevariation in the price ratio of two different consumer goodsin either the United States or Canada.

7This estimate comes from the studies by McCallum andHelliwell. Shang-Jin Wei comes up with much smaller esti-mates of home bias for the merchandise trade of a broader


23

dence indicates the home bias is apt to be evenlarger for U.S.-Canada services trade.8 Anotherstudy translates the impact of the U.S.-Cana-dian border on consumer prices into an equiva-lent physical distance, estimating that crossingthe border is equivalent to adding a distance of1780 miles between markets.9 Whether mea-sured in miles or trade volumes, the economiceffect of the U.S.-Canada border is considerable.

Two conclusions can be inferred from thisevidence. First, the U.S.-Canada border has asurprisingly large impact on both trade patternsand product prices in the two-country region.Second, if the relatively open U.S.-Canada bor-der exhibits such substantial economic effects,it is likely that borders have even greater im-pacts on trade flows and relative prices betweenthe United States and other countries. But whydo borders appear to have such large effects?

THE ECONOMIC ROLE OF THE BORDERNational borders can influence economic

activity in a number of ways. As political andlegal boundaries, they provide a means for gov-ernments to erect barriers to international flowsof goods, services, and factors of production.These measures take a variety of forms and areinstituted for a number of reasons. Tariffs drivea wedge between a domestic market and for-eign supplies, frequently with the intention ofoffering protection to domestic industries.10 Thesame is true of quotas, nontariff trade barriersthat impose quantitative restrictions on im-ports.

Other so-called nontariff barriers often havethe same effect but may or may not be erectedfor trade policy purposes. This broad categoryof barriers includes technical standards, licens-ing and certification requirements, health andsafety regulations, border formalities, and gov-ernment procurement practices. There are nu-merous instances in which regulators have beenaccused of imposing measures to protect do-mestic industries under the guise of other con-cerns such as the environment or public health.For example, in the early 1990s, Ontario levieda 10-cent tax on all beer sold in cans. The statedobjective was to encourage container re-use. ButU.S. beer manufacturers viewed the tax as pro-tectionist because, unlike Canadian beer, mostAmerican beer is sold in cans and is thus sub-ject to the levy. Practically speaking, of course,determining whether a specific nontariff bar-rier was intended to shelter domestic marketsfrom foreign competition or had some otherprimary objective is often difficult.

Other examples of government-imposedbarriers include controls on international flowsof capital and labor, limitations on holdings offoreign exchange, and market-entry and own-ership restrictions. All of these measures dif-ferentiate the products and inputs of the do-

sample of industrialized countries. However, reconcilingWei’s findings with those for U.S.-Canadian merchandisetrade is complicated by conceptual and measurement dif-ferences in the two sets of studies.

8See the paper by Helliwell and McCallum. This is likelyfor two reasons. First, the free trade agreements betweenthe U.S. and Canada did not include some service sectors,such as health, transportation, basic telecommunications,and legal services. Second, national regulations in two im-portant service sectors, broadcasting and finance, couldlimit the bilateral exchange of these services.

9See Engel and Rogers (1996). Their study covers theperiod 1978-93 while McCallum’s analysis of merchandisetrade is based on data for 1988. Because the two studiesinclude data from the period prior to the implementationof the Canada-United States Free Trade Agreement, it mightbe supposed that their estimates overstate the current im-pact of the U.S.-Canada border. However, Engel and Rogersfind that the border’s effect is no smaller when data priorto 1990 are excluded. Likewise, Helliwell’s update ofMcCallum’s work finds comparable estimates for merchan-dise trade through 1990. This could reflect the fact that theeffective trade barriers between the United States andCanada were already low before the agreement. An alter-native interpretation is that adjustment to the free tradeagreement was not complete by the early 1990s. 10Tariffs also raise revenue for the government.



mestic economy from those originating outsidethe border, effectively contributing to the estab-lishment of an economic frontier between acountry and the rest of the world.

Tariffs and other formal barriers to trade be-tween the United States and Canada have longbeen lower than in most other parts of theworld. Thus, it is unlikely that they can accountfor the lion’s share of the estimated border ef-fects. A potentially larger effect could comefrom past trade policies. High tariffs were keycomponents ofCanada’s NationalPolicy, which wasinstituted in the lat-ter part of the 19thcentury. The policysought to promoteeconomic develop-ment and east-westtransportation andtrade links withinCanada. To the ex-tent that it led to theintegration of Cana-dian markets andthe formation of strong internal distributionnetworks, this policy could bear some respon-sibility for the current home bias in Canadianmerchandise trade.11 Informal trade barriers ornontariff barriers in both countries could alsocontribute to the segmentation of U.S. and Ca-nadian markets.

The economic impact attributed to the bor-der might actually reflect the effects of geo-graphic distance between markets. If transpor-tation and information costs increase with dis-

tance, trade flows should be larger betweenmarkets that are geographically close to oneanother than between more distant markets. Forthe same reasons, price differentials across mar-kets should be smaller when the markets areclose to one another. Indeed, geographic dis-tance is a significant factor in both merchan-dise trade flows and price dispersion within theU.S.-Canada region. But the border betweenCanada and the United States appears to havea separate effect on both measures of economic

integration. Asstated earlier, tradebetween two Cana-dian provinces issubstantially greaterthan that between aprovince and anequidistant U.S.state. Moreover,even after control-ling for distance, thevariability of con-sumer prices be-tween a city inCanada and a city in

the United States is considerably higher thanthat between either two U.S. cities or two Ca-nadian cities.12

Borders are usually demarcations betweencurrency areas. Consequently, most interna-tional transactions require the exchange of onecurrency for another. Currency exchanges typi-cally entail some small cost associated withtranslating one currency into another. A smallcost for each of millions of transactions canamount to a considerable sum; one estimateplaces foreign-exchange costs in Germany at 1percent of GDP.13 However, there is a risk of

11It could be argued that Canada’s current trade pat-terns are appropriate in view of its strong internal distribu-tion networks. But a quick glance at a map suggests that,were it not for Canada’s National Policy, Ontario mighttoday have stronger links with, say, New York than withBritish Columbia.

12See Engel and Rogers’ 1996 paper.

13See “When the Walls Come Down,” The Economist, July5, 1997, pp. 61-63.

The economic impactattributed to the

border might actuallyreflect the effects ofgeographic distance

between markets.


25

substantially larger costs when a contract be-tween parties in two countries calls for futurepayment. Between the time the price is set andsettlement is made, unexpected changes in thevalue of the exchange rate will alter the ulti-mate price of the transaction for one of the par-ties involved. Unlike exchanges within a singlecountry or currency area, international trans-actions often entail exchange-rate risk. This riskcould act as a barrier to international trade. Inempirical studies of international trade, cur-rency risk is commonly measured by the vola-tility of the relevant exchange rate. However,perhaps because financial instruments such asforward exchange contracts are available to re-duce or eliminate currency risk, such studieshave yielded mixed results, and there is cur-rently no consensus among economists thatexchange-rate volatility has had a significantnegative impact on trade volumes.

When price comparisons are used to mea-sure border effects, the exchange rate mattersin a different way. International price compari-sons are made by using the nominal exchangerate to translate prices into a common currency.However, the nominal exchange rate is typicallymore variable than product prices. By implica-tion, much of the variation in the relative dol-lar prices of Canadian and U.S. consumer prod-ucts could simply reflect fluctuations in thenominal exchange rate between the two coun-tries. Indeed, the empirical evidence indicatesthat changes in the exchange rate are signifi-cant factors in the volatility of relative U.S.-Ca-nadian prices. But they are far from the wholestory.14

Several economists have noted that consum-

ers exhibit a distinct home bias, preferring todeal with firms in their own country and topurchase domestic products. Little is knownabout the precise reasons for this preference,but a number of factors may be involved. Tothe extent that they define social boundaries,national borders may also represent the eco-nomic effects of distinct tastes, history, tradi-tions, and cultures. Alternatively, a preferencefor home products may simply reflect ignoranceabout or lack of access to alternatives. Regionswithin a common border typically share net-works of associations, as well as legal, finan-cial, and regulatory systems. Not only can thisease the acquisition of information but, onceobtained, such knowledge is often universallyapplicable within the border. In addition, mar-keting and distribution networks for goods,services, and inputs may be more integratedwithin each country than they are across bor-ders.15 These networks may make it easier tolearn about and gain access to domestic prod-ucts, contributing to a home bias. Although itis difficult to measure the contribution of thesefactors to the economic role of the border, theyshould not be dismissed as necessarily trivial.

CONCLUSIONDespite evidence that the U.S. economy has

become more open, recent empirical researchfinds that the border between the United Statesand Canada has a very large impact on bilat-eral trade flows and relative prices. Given therelative openness of the U.S.-Canada border, itis unlikely that the border’s effects are any lesssignificant between product markets in theUnited States and other countries. This contra-dicts the notion that globalization has alreadyrendered national borders economically mean-ingless. But because most of the evidence isbased on relatively recent data, it is not known14Engel and Rogers (1996) explore the possibility that

the effect attributed to the U.S.-Canada border is, in fact,the product of fluctuations in nominal exchange rates andrigidity in local prices. They find that while local price ri-gidity is responsible for part of the measured border effect,it accounts for less than half of it.

15See the 1995 paper by Engel and Rogers for a modelof international trade with marketing costs.



whether the border’s economic impacts are ac-tually smaller now than in the past.

The reasons for the border’s substantial ef-fects are not yet completely understood. Con-sequently, it is difficult to speculate how recentadvances in communication like the Internetwill ultimately reduce the economic boundaries

between nations. However, the effects of theborder appear to extend beyond the economicimpacts of geographic distance and formaltrade barriers. By implication, merely liberaliz-ing trade or reducing transportation costs be-tween national markets may not be enough tocause the border to disappear.


27

References

Engel, Charles. “Real Exchange Rates and Relative Prices,” Journal of Monetary Economics, 32 (1993),pp. 35-50.

Engel, Charles, and John Rogers. “Regional Patterns in the Law of One Price: The Roles of Geographyvs Currencies,” Working Paper 5395, National Bureau of Economic Research (1995).

Engel, Charles, and John Rogers. “How Wide Is the Border?” American Economic Review, 86 (Decem-ber, 1996), pp. 1112-25.

Feldstein, Martin. “Tax Policy and International Capital Flows,” Working Paper 4851, National Bu-reau of Economic Research (1994).

French, Kenneth, and James Poterba. “Investor Diversification and International Equity Markets,”American Economic Review, 81 (May, 1991), pp. 222-26.

Helliwell, John. “Do National Borders Matter for Quebec’s Trade?” Canadian Journal of Economics, 29(August, 1996), pp. 507-22.

Helliwell, John, and John McCallum. “National Borders Still Matter for Trade,” Policy Options, 13(July/August, 1995), pp. 44-48.

Krugman, Paul, and Robert Z. Lawrence. “Trade, Jobs, and Wages,” Scientific American (April, 1994),pp. 44-49.

McCallum, John. “National Borders Matter: Canada-U.S. Regional Trade Patterns,” American Eco-nomic Review, 85 (June, 1995), pp. 615-23.

Ohmae, Kenichi. The Borderless World. New York: Harper Business, 1990.

Rauch, James. “Networks Versus Markets in International Trade,” Working Paper 5617, National Bu-reau of Economic Research (1996).

Rogers, John, and Michael Jenkins. “Haircuts or Hysteresis? Sources of Movements in Real ExchangeRates,” Journal of International Economics, 38 (1995), pp. 339-60.

Tesar, Linda, and Ingrid Werner. “Home Bias and the Globalization of Securities Markets,” WorkingPaper 4218, National Bureau of Economic Research (1992).

Wei, Shang-Jin. “Intra-national versus International Trade: How Stubborn Are Nations in Global Inte-gration?” Working Paper 5531, National Bureau of Economic Research (1996).

Documents

Does Trade with Low-Wage Countries Hurt American Workers? · Does Trade with Low-Wage Countries Hurt American Workers? Stephen Golub* T here are gaping disparities in wages and benefits